The purpose of this project is to construct up-to-date linkage maps of the human chromosomes based on all relevant studies, without access to the raw phenotype data, and to test hypotheses concerning the chiasma map, interference, race and sex differences, and disease heterogeneity. State of the art statistical methods will be applied to updating linkage information on pairwise recombination fractions and on gene order, determined from multilocus analyses, in the construction of the linkage maps . Empirical Bayes' methods (parametric and nonparametric) will be developed, programmed for computer implementation, and compared with existing likelihood and pure Bayesian approaches, by cross- validation and computer simulation. The linkage maps constructed will be compared with those estimated by maximum likelihood methods from pairwise lod score data, and information from confirmed physical orders of loci will be included in the analyses. In addition, all published human linkage data will be synthesized, and after thorough checking, this linkage data base will be accessible via a computer terminal and modem to all interested researchers and clinicians through the New Haven Human Gene Mapping Library. At present this data base contains lod score data for pairs of markers obtained from 35OO references, and updating it will be a continuous process. It is essential that the linkage maps and the data set used to construct them be kept up to date, and be readily accessible as a secure base for formal genetics and genetic counseling. All applications of linkage to antenatal diagnosis and genetic counseling depend on the frequency with which linked genes are transmitted together to progeny. Precise chromosome maps based on recombination rates are required for determining these transmission probabilities; they cannot be inferred from physical distances. Thus accurate linkage (recombination) maps are essential for applying linkage to genetic counseling.